Recently I’ve got interesting task related to random data generation and writing this data on disk as fast as possible using Java. Unfortunately performance requirements were really tight: in peak application should support writing at ~3 Gb/s. Which means that it should generate random data at really high pace, encode it to proper charset and finally write to disk. That’s pretty much an exposition and motivation behind following measurements and why it was so important to dig in.

Recently I’ve bought both Raspberry Pi and Tinker Board to experiment with microservices performance, cause in the cloud you never know how busy your host is. As an application under test I’ve found just perfect match - Dockerised Microservices Calculator by Jeremy Cook written on node.js. It’s dockerised microservices calculator that has been designed to evaluate simple and complex mathematical expressions with any valid user provided expression using the order of operations…